Why don't we close this ticket? http://trac.sagemath.org/sage_trac/ticket/4566
There seems to be already a quite reasonable explanation to do this in its comments Maurizio On 4 Set, 15:34, Burcin Erocal <bur...@erocal.org> wrote: > Hi Jason and anybody interested in symbolics in Sage, > > On Fri, 04 Sep 2009 07:33:12 -0500 > > > > Jason Grout <jason-s...@creativetrax.com> wrote: > > > Golam Mortuza Hossain wrote: > > > > On Sep 3, 10:52 am, Nicolas <nicolas.fresseng...@gmail.com> wrote: > > >> I am trying to do the following thing but I am not sure whether > > >> this is possible in sage (though I think it is in Mathematica) > > > >> Let's say I define a symbolic function > > > >> sage: f=function('f',x) > > > >> I would like to issue some command to "tell" sage what the > > >> derivative of f is (for instance : g) so that the command > > > >> sage: diff(f) > > > >> would yield g > > > > Yes, its possible also in Sage. I am giving you an example that you > > > can adapt for your need. (if you have have a function of multiple > > > arguments then you should read diff_param keywords. For example, > > > try heaviside?? in sage prompt to see how to read diff_param) > > > > -------- > > > sage: g = function('g') > > > sage: def f_deriv(*args, **kwds): > > > ....: return g(*args) > > > ....: > > > sage: f = function('f', derivative_func=f_deriv) > > > Where is this documented? I see some examples (of this and lots of > > other keywords) under: > > > sage.symbolic.function.SFunction?? > > > But I don't think most users would use sage_search or grep to find it > > like I had to resort to after guessing a bunch of places the > > documentation might be... > > > It looks like there are quite a lot of very interesting and useful > > arguments to function. I had no idea these existed! > > This is some of the new functionality that came with the GiNaC wrapper. > It was one of the first things I wrapped, and in the rush for the > switch over from maxima, it still remains in its raw state. > > The interface to define new functions needs more care and polish. ATM, > you define a class with methods _eval_, _evalf_, _derivative_, etc. and > your function ends up being an instance of this class. I find this > rather confusing and would appreciate any suggestions on other design > options. > > There are still lots of rough edges in the new symbolics code. I don't > have time to fix all these problems or write documentation. Though, I > will do my best to answer questions and explain any tricky parts of the > new symbolics to anybody interested, and put in the work to make the > required changes (if any) at the c++ level in pynac. > > So if you're interested in getting the symbolics up to snuff, please > take the plunge and look at the new symbolics code in sage/symbolic/. > > There are lot's of symbolics bugs on trac: > > http://is.gd/2SqFy[symbolics]http://is.gd/2SqMs[calculus] > > I can help fix the ones related to new symbolics, #6220 looks like a > good start for example. > > If there is interest we can even hold a symbolics bug day where we meet > on IRC and work on some of these. > > Cheers, > Burcin > > P.S. Many thanks to Karl-Dieter and Golam for the patches waiting for > review on trac. I'll try to do the reviews this weekend if no one gets > there first. --~--~---------~--~----~------------~-------~--~----~ To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---